U.S. patent application number 14/459206 was filed with the patent office on 2014-11-27 for content recommendation system, content recommendation method, content recommendation device, and information storage medium.
The applicant listed for this patent is Sony Corporation. Invention is credited to Nozomu Ikeda, Kazuo Ishii, Masaaki Isozaki, Wataru Onogi, Katsu Saito, Yoshikazu Takahashi.
Application Number | 20140351207 14/459206 |
Document ID | / |
Family ID | 43449240 |
Filed Date | 2014-11-27 |
United States Patent
Application |
20140351207 |
Kind Code |
A1 |
Saito; Katsu ; et
al. |
November 27, 2014 |
CONTENT RECOMMENDATION SYSTEM, CONTENT RECOMMENDATION METHOD,
CONTENT RECOMMENDATION DEVICE, AND INFORMATION STORAGE MEDIUM
Abstract
The content recommendation system includes a user attribute
information acquisition means for acquiring attribute information
of a given user, a list acquisition means for acquiring at least
one list from among lists of content sequentially generated over
time on the basis of the attribute information of the given user
acquired by the user attribute information acquisition means, a
user preference information acquisition means for acquiring user
preference information, which is feature information of content
preferred by the given user, a content extraction means for
extracting some content from content included in the list acquired
by the list acquisition means on the basis of the user preference
information acquired by the user preference information acquisition
means, and a content presentation means for presenting the content
extracted by the content extraction means to the given user.
Inventors: |
Saito; Katsu; (Saitama,
JP) ; Isozaki; Masaaki; (Kanagawa, JP) ;
Onogi; Wataru; (Kanagawa, JP) ; Ishii; Kazuo;
(Tokyo, JP) ; Ikeda; Nozomu; (Tokyo, JP) ;
Takahashi; Yoshikazu; (Saitama, JP) |
|
Applicant: |
Name |
City |
State |
Country |
Type |
Sony Corporation |
Tokyo |
|
JP |
|
|
Family ID: |
43449240 |
Appl. No.: |
14/459206 |
Filed: |
August 13, 2014 |
Related U.S. Patent Documents
|
|
|
|
|
|
Application
Number |
Filing Date |
Patent Number |
|
|
13380528 |
Dec 23, 2011 |
8838641 |
|
|
PCT/JP2010/059678 |
Jun 8, 2010 |
|
|
|
14459206 |
|
|
|
|
Current U.S.
Class: |
707/609 ;
707/748 |
Current CPC
Class: |
G06F 16/9535 20190101;
G06F 16/435 20190101; G06F 16/24578 20190101; G06F 16/4387
20190101 |
Class at
Publication: |
707/609 ;
707/748 |
International
Class: |
G06F 17/30 20060101
G06F017/30 |
Foreign Application Data
Date |
Code |
Application Number |
Jul 14, 2009 |
JP |
2009-165391 |
Claims
1-9. (canceled)
10. An apparatus for providing content data, comprising: a ranking
unit that ranks a plurality of content data according to user
browsing information using a processor; and a selection unit that
extracts a part of the plurality of content data based on the
ranking and a user preference of content data.
11. The apparatus of claim 10, the ranking unit comprising: an
aggregator that receives stored user attributes and the user
browsing information and generates aggregated user browsing
information for combinations of content data ids and user attribute
categories, wherein the content data ids are associated with the
plurality of content data; and an internal rankings generation unit
that periodically generates rankings for subsets of the content
data ids corresponding to user attribute categories based on
changes in the aggregated user browsing information.
12. The apparatus of claim 11, the selection unit comprising: a
ranking selection unit that receives transmitted user attributes
and generates a synthesized list of content data ids based on the
ranked subsets that correspond to user attribute categories
encompassing the transmitted user attributes; and a secondary list
generation unit that filters and sorts the synthesized list of
content data ids based on a user preference of content data and
provides the sorted filtered synthesized list of content data ids
to a user device for retrieving the content data associated with
the content data ids.
13. The apparatus of claim 11, wherein user browsing information
includes indications of user access to text related to content
data.
14. The apparatus of claim 12, wherein the user preference of
content data comprises content attribute information.
15. The apparatus of claim 14, wherein the user preference of
content data is received from a user device in response to a user
designation of the content attribute information.
16. The apparatus of claim 12, wherein the user preference of
content data comprises a user preference vector.
17. The apparatus of claim 16, further comprising updating the user
preference vector based on content data evaluated by the user.
18. The apparatus of claim 11, wherein stored user attributes
include age, residence, and language, and attribute categories
include combinations of age range, residence, and language.
19. The apparatus of claim 13, wherein a change in aggregated user
browsing information depends on an increase in a number of user
access indications for a combination of a content data id and a
user attribute category.
20. The apparatus of claim 14, wherein filtering the synthesized
list of content data ids based on a user preference of content data
comprises receiving content attribute information for content data
ids in the synthesized list of content data ids from a content
attribute database and comparing them to the content attribute
information of the user preference of content data.
21. The apparatus of claim 16, wherein the secondary list
generation unit sorts the filtered synthesized list of content data
ids based on the user preference vector and feature vectors
corresponding to the content data ids of the filtered synthesized
list of content data ids, the values for elements of the feature
vectors automatically determined from the content ids for the
feature vectors.
22. The apparatus of claim 21, further comprising incrementing the
user preference vector towards a feature vector for a content data
id, based on a received positive evaluation of the content data
associated with the content data id; and wherein the user
preference vector is stored by the system.
23. The apparatus of claim 21, wherein the secondary list
generation unit sorts the filtered synthesized list of content data
ids in increasing order of the angle between the user preference
vector and the feature vectors for the data content ids in the
filtered synthesized list of content data ids.
24. The apparatus of claim 12 in which the sorted filtered
synthesized list of content data ids is provided by a different
content distributor than the content data.
25. A method for providing content data, the method comprising:
ranking a plurality of content data according to user browsing
information using a processor; and extracting a part of the
plurality of content data based on the ranking and a user
preference of content data.
26. The method of claim 25, wherein ranking a plurality of content
data comprises: receiving at a communication interface stored user
attributes and user browsing information; aggregating user browsing
information for combinations of content data ids and user attribute
categories using a processor, wherein the content data ids are
associated with a plurality of content data; and generating
rankings for subsets of the content data ids corresponding to user
attribute categories using the processor based on changes in the
aggregated user browsing information.
27. The method according to claim 26, wherein extracting a part of
the plurality of content data based on the ranking and a user
preference of content data comprises: receiving at the
communication interface transmitted user attributes and a user
preference of content data; generating a sorted list of content
data ids based on the rankings for subsets, the transmitted user
attributes, and the user preference of content data; and providing
the sorted filtered synthesized list of content data ids to a user
device for use in retrieving the content data associated with the
content data ids.
28. The method of claim 25, wherein the user browsing information
includes indications of user access to text related to content
data.
29. The method of claim 25, wherein the user preference of content
data comprises content attribute information.
30. The method of claim 25, wherein the user preference of content
data comprises a user preference vector.
31. The method of claim 30, further comprising updating the user
preference vector based on content data evaluated by the user.
32. The method of claim 29, wherein the user preference of content
data is received from a user device in response to a user
designation of the content attribute information.
33. A non-transitory computer-readable medium storing instructions
that, when executed by a processor, cause a computer to perform the
operations comprising: ranking a plurality of content data
according to user browsing information using a processor; and
extracting a part of the plurality of content data based on the
ranking and a user preference of content data.
34. The non-transitory computer-readable medium of claim 33,
wherein ranking a plurality of content data comprises: receiving at
a communication interface stored user attributes and user browsing
information; aggregating user browsing information for combinations
of content data ids and user attribute categories using a
processor, wherein the content data ids are associated with a
plurality of content data; and generating rankings for subsets of
the content data ids corresponding to user attribute categories
using the processor based on changes in the aggregated user
browsing information.
35. The non-transitory computer-readable medium of claim 34,
wherein extracting a part of the plurality of content data based on
the ranking and a user preference of content data comprises:
receiving at the communication interface transmitted user
attributes and a user preference of content data; generating a
sorted list of content data ids based on the rankings for subsets,
the transmitted user attributes, and the user preference of content
data; and providing the sorted filtered synthesized list of content
data ids to a user device for use in retrieving the content data
associated with the content data ids.
36. The non-transitory computer-readable medium of claim 33,
wherein the user browsing information includes indications of user
access to text related to content data.
37. The non-transitory computer-readable medium of claim 33,
wherein the user preference of content data comprises content
attribute information.
38. The non-transitory computer-readable medium of claim 33,
wherein the user preference of content data comprises a user
preference vector.
39. The non-transitory computer-readable medium of claim 38,
further comprising updating the user preference vector based on
content data evaluated by the user.
40. The non-transitory computer-readable medium of claim 37,
wherein the user preference of content data is received from a user
device in response to a user designation of the content attribute
information.
Description
TECHNICAL FIELD
[0001] The present invention relates to a content recommendation
system, a content recommendation method, a content recommendation
device, and an information storage medium, and more particularly,
to content recommendation technology.
BACKGROUND ART
[0002] Recently, it has become possible to enjoy what one desires
from among a great number of pieces of content by use of a
communication network such as the Internet. Various recommendation
technologies have been proposed according to a great number of
pieces of available content. For example, in content-based
filtering, content to be recommended to a user among a number of
pieces of content is extracted according to similarity between
feature information of content and feature information of content
preferred by the user.
SUMMARY OF INVENTION
Technical Problem
[0003] However, the content-based filtering has a problem in that
the same content is continuously recommended to the user as long as
the feature information of the content preferred by the user does
not change.
[0004] The present invention has been made in view of the
above-described problem, and an object of the invention is to
provide a content recommendation system, a content recommendation
method, a content recommendation device, and an information storage
medium that can diversify content to be recommended to a user.
Solution to Problem
[0005] According to the present invention in order to solve the
above-described problem, there is provided a content recommendation
system characterized in including: a user attribute information
acquisition means for acquiring attribute information of a given
user; a list acquisition means for acquiring at least one list from
among lists of content sequentially generated over time on the
basis of the attribute information of the given user acquired by
the user attribute information acquisition means; a user preference
information acquisition means for acquiring user preference
information, which is feature information of content preferred by
the given user; a content extraction means for extracting some
content from content included in the list acquired by the list
acquisition means on the basis of the user preference information
acquired by the user preference information acquisition means; and
a content presentation means for presenting the content extracted
by the content extraction means to the given user.
[0006] According to the present invention, there is provided a
content recommendation method characterized in including: a user
attribute information acquisition step of acquiring attribute
information of a given user; a list acquisition step of acquiring
at least one list from among lists of content sequentially
generated over time; a user preference information acquisition step
of acquiring user preference information, which is feature
information of content preferred by the given user; a content
extraction step of extracting some content from content included in
the list acquired in the list acquisition step on the basis of the
user preference information acquired in the user preference
information acquisition step; and a content presentation step of
presenting the content extracted in the content extraction step to
the given user.
[0007] According to the present invention, there is provided a
content recommendation device characterized in including: a user
attribute information acquisition means for acquiring attribute
information of a given user; a list acquisition means for acquiring
at least one list from among lists of content sequentially
generated over time; a user preference information acquisition
means for acquiring user preference information, which is feature
information of content preferred by the given user; a content
extraction means for extracting some content from content included
in the list acquired by the list acquisition means on the basis of
the user preference information acquired by the user preference
information acquisition means; and a content presentation means for
presenting the content extracted by the content extraction means to
the given user.
[0008] According to the present invention, there is provided a
program for causing a computer to function as: a user attribute
information acquisition means for acquiring attribute information
of a given user; a list acquisition means for acquiring at least
one list from among lists of content sequentially generated over
time; a user preference information acquisition means for acquiring
user preference information, which is feature information of
content preferred by the given user; a content extraction means for
extracting some content from content included in the list acquired
by the list acquisition means on the basis of the user preference
information acquired by the user preference information acquisition
means; and a content presentation means for presenting the content
extracted by the content extraction means to the given user. An
information storage medium according to the present invention
stores the above-described program.
BRIEF DESCRIPTION OF DRAWINGS
[0009] FIG. 1 is an entire configuration diagram of a content
recommendation system according to an embodiment of the present
invention.
[0010] FIG. 2 is a hardware configuration diagram of a server.
[0011] FIG. 3 is a hardware configuration diagram of a user
device.
[0012] FIG. 4 is an external perspective diagram of the user
device.
[0013] FIG. 5 is an external perspective diagram of a user device
according to a modified example.
[0014] FIG. 6 is a functional block diagram of the user device.
[0015] FIG. 7 is a functional block diagram of a music distribution
server.
[0016] FIG. 8 is a diagram showing storage content of a user
information storage unit.
[0017] FIG. 9 is a configuration diagram of an aggregate database
(DB).
[0018] FIG. 10 is a diagram showing storage content of an internal
ranking storage unit.
[0019] FIG. 11 is a diagram showing ranking synthesis.
[0020] FIG. 12 is a configuration diagram of a music feature
DB.
[0021] FIG. 13 is a configuration diagram of a music attribute
DB.
[0022] FIG. 14 is an operation sequence diagram of the content
recommendation system according to an embodiment of the present
invention.
DESCRIPTION OF EMBODIMENTS
[0023] Hereinafter, preferred embodiments of the present invention
will be described in detail with reference to the appended
drawings. Note that, in this specification and the drawings,
elements that have substantially the same function and structure
are denoted with the same reference signs, and repeated explanation
is omitted.
[0024] FIG. 1 is an entire configuration diagram of a content
recommendation system according to an embodiment of the present
invention. As shown in the same drawing, the content recommendation
system 10 is connected to a data communication network 18 such as
the Internet, and includes a music distribution server 14, a music
ranking distribution server 15, and a plurality of user devices 12,
which can perform data communication with each other. The user
device 12 is, for example, a computer system such as a personal
computer, a computer game system, or a home server installed at
home or a portable computer system such as a portable game device
or a mobile phone, and receives a list of music to be recommended
to a user of the user device 12 by accessing the music distribution
server 14. The user device 12 requests the music distribution
server 14 to provide music data included in the list, and receives
and reproduces the music data. On the other hand, the music
distribution server 14 is constituted, for example, by a computer
system such as a well-known server computer, and transmits a list
of music to be recommended to a user of each user device 12 to the
user device 12. Individual music data is transmitted in response to
a request from each user device 12. Here, although an application
example of the present invention for a music recommendation has
been described, the present invention is not limited thereto. Of
course, it is possible to apply the present invention to recommend
various content, for example, moving images such as movies, still
images such as photos, documents such as novels, and the like. The
music ranking distribution server 15 is constituted, for example,
by a computer system such as a well-known server computer managed
by a person different from a manager of the music distribution
server 14, and returns music rankings in response to a request from
the music distribution server 14. The music rankings are issued,
for example, periodically such as weekly or monthly, for example,
according to a music genre such as pop, jazz, or classic, by
nationality, and each ranking is stored in the music distribution
server 14 in association with an issuance time and a music genre.
Rankings may be created in various viewpoints, and may be, for
example, music rankings according to an increase in the number of
purchases, music rankings according to an increase in the number of
downloads, music rankings according to an increase in the number of
times information related to music (for example, supporting
documentation) is browsed, or the like.
[0025] FIG. 2 is a diagram showing a hardware configuration example
of the music distribution server 14 and the music ranking
distribution server 15. As shown in the same drawing, the music
distribution server 14 or the music ranking distribution server 15
includes a processor 20, a memory 21, a hard disk drive (HDD) 23, a
medium drive 24, and a communication interface (I/F) 26, which are
connected to a bus 22 and configured to exchange data with one
another. The memory 21 is constituted by a read only memory (ROM)
and a random access memory (RAM). The ROM stores various system
programs, and the RAM is mainly used to work for the processor 20.
In the HDD 23, a program for music distribution or music ranking
distribution is stored, or various DBs for the music distribution
or music ranking distribution are constructed. The medium drive 24
is a device that reads data stored in a computer-readable medium 25
such as a compact disk-read only memory (CD-ROM) or a digital
versatile disk-random access memory (DVD-RAM), or writes data
thereto. The communication I/F 26 controls an operation in which
the music distribution server 14 or the music ranking distribution
server 15 performs data communication with another computer system
via the communication network 18. The processor 20 controls each
part of the server according to a program stored in the memory 21,
the HDD 23, or the medium 25.
[0026] On the other hand, FIG. 3 is a diagram showing a hardware
configuration example of the user device 12. As shown in the same
drawing, the user device 12 includes a processor 31, a memory 32, a
display control unit 33, a sound control unit 34, an HDD 35,
manipulation devices 36, a medium drive 37, and a communication I/F
38, which are connected to a bus 39 and configured to exchange data
with one another. The memory 32 is constituted by a ROM and a RAM.
The ROM stores various system programs, and the RAM is mainly used
to work for the processor 31. In the HDD 35, various programs such
as a music reproduction program are stored, or various DBs are
constructed. The medium drive 37 is a device that reads data stored
in a computer-readable medium 40 such as a CD-ROM or a DVD-RAM, or
writes data thereto. The communication I/F 38 controls an operation
in which the user device 12 performs data communication with
another computer system such as the music distribution server 14
via the communication network 18. The display control unit 33 is
configured to include a video memory, and converts an image drawn
in the video memory by the processor 31 into a video signal,
outputs the video signal to a display, and causes the image to be
displayed. The sound control unit 34 is configured to include a
sound buffer, and the processor 31 converts sound data stored in
the sound buffer into an analog sound signal and outputs the analog
sound signal to speakers to be audibly outputted. The manipulation
devices 36 are means for allowing the user to provide an
instruction to the user device 12, and are, for example, a
keyboard, a pointing device such as a mouse, a gamepad, or the
like. Furthermore, the program 31 controls each part of the server
device 12 according to a program stored in the memory 32, the HDD
35 or the medium 40.
[0027] The user device 12 may be implemented in various types, and
configured, for example, as a stationary game console operable by
household power as shown in FIG. 4. In this case, each hardware
element shown in FIG. 3 is accommodated in a housing 43, and a
display device 41a and built-in speakers 42 and 42 of a television
receiver 41 separate from the housing 43 are used as the display
and the speakers. The manipulation devices 36 are also provided
separately from the housing 43.
[0028] Alternatively, the user device 12 may be configured as a
portable and integrated game device operable by a battery shown in
FIG. 5. In this case, each hardware element shown in FIG. 3 is
accommodated in a housing 44, and furthermore a flat panel display
45 provided on a surface of the housing 44 is used as the display.
The manipulation devices 36 are also provided on the surface of the
housing 44, and are arranged, for example, on the left and right of
the flat panel display 45. Speakers (not shown) embedded in the
housing 44 or stereo headphones 46 separate from the housing 44 may
be used as the speakers.
[0029] Here, a functional configuration of the user device 12 will
be described. FIG. 6 is a functional block diagram of the user
device 12. As shown in the same drawing, the user device 12 is
configured to functionally include a manipulation unit 61 and a
music reproduction unit 62. The above-described functional elements
are implemented by a program executed in the user device 12. The
manipulation unit 61 is configured around the manipulation devices
36, and transmits a request of a music list (including a user ID)
to the music distribution server 14 via the communication I/F 38 if
a specific request manipulation to the manipulation devices 36 is
performed. In the request, a user ID and a designation of an
attribute of music (for example, a mood such as relaxed, ballad,
happy, or active) are included. The attribute of music may be
designated by the user using the manipulation devices 36, or may be
randomly selected. A preference vector to be described later may be
managed by the user device 12. In this case, user preference
information managed may be included in the request of the music
list. The music reproduction unit 62 to be described later receives
a music list returned by the music distribution server 14 via the
communication network 18 in response to the request, and transmits
identification information of each piece of music included in the
music list, that is, a music ID, to the music distribution server
14 via the communication I/F 38. The music reproduction unit 62
reproduces music data returned from the music distribution server
14 using the sound control unit 34 in response to the transmission
of the music ID. At this time, the music reproduction unit 62
displays song title data included in the music data on the display
as shown in FIGS. 4 and 5.
[0030] When an explicit negative manipulation (a specific
manipulation indicating that the user does not like the
corresponding music) to the manipulation devices 36 is performed
during music reproduction, the manipulation unit 61 transmits a
negative evaluation to the music distribution server 14 along with
a music ID of reproduced music and a user ID. When a skip
manipulation to the manipulation devices 36 is performed during
music reproduction, the music reproduction unit 62 is notified of
the skip manipulation. The music reproduction unit 62 stops the
reproduction of music according to the notification, transmits the
next music ID to the music distribution server 14, and reproduces
returned music data. Even when the skip manipulation is performed
as described above, the manipulation unit 61 may be configured to
transmit a negative evaluation to the music distribution server 14
along with a music ID of reproduced music and a user ID. When music
is fully reproduced without being skipped, the music reproduction
unit 62 notifies the manipulation unit 61 of the fact that the
music has been fully reproduced without being skipped. In this
case, the manipulation unit 61 transmits a positive evaluation to
the music distribution server 14 along with a music ID of fully
reproduced music and a user ID. Furthermore, when an explicit
positive manipulation (a specific manipulation indicating that the
user likes the corresponding music) to the manipulation devices 36
is performed during music reproduction, a positive evaluation is
transmitted to the music distribution server 14 along with a music
ID of reproduced music and a user ID. The positive or negative
evaluation generated for each piece of music delivered to the music
distribution server 14 is used to generate a preference vector
indicating a feature of music preferred by each user or generate
rankings of pieces of music preferred by a plurality of users in
the music distribution server 14.
[0031] Next, a functional configuration of the music distribution
server 14 will be described. FIG. 7 is a functional block diagram
of the music distribution server 14. As shown in the same drawing,
the music distribution server 14 is configured to functionally
include a user information storage unit 71, an aggregate unit 72,
an internal ranking generation unit 73, an internal ranking storage
unit 74, a ranking selection and synthesis unit 75, a preference
vector update unit 77, a preference vector storage unit 78, a
sorting unit 79, a secondary list generation unit 80, a primary
list storage unit 81, a music attribute DB 82, a music feature DB
83, a music distribution unit 84, and a music DB 85. These
functional elements are implemented by a program executed in the
music distribution server 14.
[0032] The user information storage unit 71 is configured using the
HDD 23 or a separate DB (not shown), and stores attributes of
users. FIG. 8 schematically shows storage content of the user
information storage unit 71. User IDs, which are identification
information of users, and attributes of the users such as age, a
city of residence, and use language are associated and stored. The
storage content of the user information storage unit 71 may be
pre-registered from the user device 12.
[0033] The aggregate unit 72 aggregates information related to an
evaluation transmitted from the user device 12 and information
related to a music list transmitted to the user device 12. The
aggregate unit 72 includes an aggregate DB 72a configured using the
HDD 23 or a separate DB (not shown). An aggregate result is stored
in the aggregate DB 72a. FIG. 9 schematically shows storage content
of the aggregate DB 72a in which a music ID, which is music
identification information, and an aggregate value x-y-z of a user
attribute range (in which age is 13 years old or less, a city of
residence is Tokyo, use language is Japanese, and the like) are
associated and stored. Here, the aggregate value includes three
numeric values: the number of times x included in a music list
transmitted to the user device 12, the number of times y a positive
evaluation for corresponding music is transmitted from the user
device 12, and the number of times z a negative evaluation for
corresponding music is transmitted from the user device 12. If a
music ID, a user ID, and a positive or negative evaluation are
received from the user device 12, the aggregate unit 72 reads
attributes of a user associated with the user ID received from the
user information storage unit 71, and updates the number of times y
or the number of times z in an aggregate value of a user attribute
range to which the attributes of the user belong. Specifically, the
number of times y is incremented by 1 if the positive evaluation is
received, and the number of times z is incremented by 1 if the
negative evaluation is received. When the sorting unit 79 generates
a music list and transmits the generated music list to the user
device 12, a music ID and a user ID of a user of the user device 12
of a destination of the music list included in the music list are
acquired, and attributes of the user associated with the user ID
are read from the user information storage unit 71. In the
aggregate value of the user attribute range to which the read
attributes belong, the number of times x described above is
incremented by 1.
[0034] The internal ranking generation unit 73 generates a ranking
for a range of various user attributes periodically such as weekly
or monthly on the basis of storage content of the aggregate DB 72a,
and stores the ranking in the internal ranking storage unit 74. As
shown in FIG. 10, the internal ranking storage unit 74 is
configured using the HDD 23 or a separate DB (not shown), and
various rankings generated by the internal ranking generation unit
73 are stored in association with generation times and user
attribute ranges thereof. For example, rankings of pieces of music
preferred by a user who is 15 years old or less, whose city of
residence is Tokyo, and whose use language is Japanese are
generated by listing IDs of a predetermined number of (for example,
100) pieces of music in descending order of total values of the
number of times y recorded in columns of "13 years old or
less/Tokyo/Japanese," "14 years old/Tokyo/Japanese," and "15 years
old/Tokyo/Japanese" shown in FIG. 9. At this time, the rankings may
be generated by listing a predetermined number of music IDs in
descending order of a ratio of a total value of the number of times
y to a total value of the number of times x described above, that
is, a ratio of the number of positive evaluations to the number of
recommendations.
[0035] The ranking selection and synthesis unit 75 reads the
attributes of the user associated with the user ID included in the
request of the music list transmitted from the user device 12 from
the user information storage unit 71, and reads rankings associated
with a range including the attributes of the user from the internal
ranking storage unit 74. The rankings corresponding to the
attributes of the user are received from the music ranking
distribution server 15. For example, the latest rankings in the
city of residence of the user may be received, and the rankings
issued in the city of residence of the user when the user was 15
years old may be received on the basis of the age of the user. The
ranking selection and synthesis unit 75 receives the rankings
associated with the attributes of the user as described above and
synthesizes the rankings as shown in FIG. 11. At this time, all
music IDs included in rankings may not be included in a list after
synthesis.
[0036] The ranking selection and synthesis unit 75 may read the
rankings associated with the range including the attributes of the
user and rankings associated with a range adjacent thereto from the
internal ranking storage unit 74, and synthesize the rankings. In
this case, the number of pieces of music included in the list after
the synthesis from the rankings associated with the range including
the attributes of the user may be less than the number of pieces of
music included in the list after the synthesis from the rankings
associated with the adjacent range. The rankings synthesized by the
ranking selection and synthesis unit 75, that is, a list of music
IDs, are stored in the primary list storage unit 81 configured
using the HDD 23 or a separate DB (not shown).
[0037] Next, the music feature DB 83 is configured using the HDD 23
or a separate DB (not shown) in which music IDs and a plurality of
feature quantities are associated and stored as shown in FIG. 12.
The features are, for example, a music tempo, a degree of sound of
a specific frequency included in music, a frequency of a specific
keyword included in expository text of music, and the like. The
feature quantities may be input as results of analysis processing
by a computer. Hereinafter, a vector having each feature quantity
as a component is referred to as a feature vector.
[0038] The preference vector storage unit 78 is configured using
the HDD 23 or a separate DB (not shown), and stores a preference
vector of each user. Like the feature vector of each piece of
music, the preference vector is a vector having each feature
quantity as a component shown in FIG. 12. The preference vector may
be generated on the basis of a feature vector of music preferred by
the user.
[0039] The preference vector update unit 77 updates a preference
vector stored in the preference vector update unit 77 on the basis
of information related to an evaluation for each piece of music
transmitted from the user device 12. For example, if a music ID, a
user ID, and a positive evaluation are received from the user
device 12, a feature vector associated with the received music ID
is read from the music feature DB 83, and a direction of the
preference vector stored in the preference vector storage unit 78
associated with the received user ID is set to be close to a
direction of the read feature vector. The preference vector update
unit 77 is provided in the side of the user device 12, so that the
preference vector update unit 77 within the user device 12 can
update the preference vector stored in the user device 12 on the
basis of information related to an evaluation for each piece of
music obtained by the manipulation unit 61. At this time, the
preference vector update unit 77 may download a feature vector of
music from the music feature DB 83 within the music distribution
server 14, or a DB equal to the music feature DB 83 is provided in
the user device 12, so that the feature vector of music can be read
from the DB. Thereby, a preference vector constantly managed within
the user device 12 to be most recent may be included in a request
of the music list as described above and transmitted to the music
distribution server 14. The music distribution server 14 receiving
the preference vector stores the preference vector in the
preference vector storage unit 78 of the music distribution server
14.
[0040] The music attribute DB 82 is configured using the HDD 23 or
a separate DB (not shown) in which a music ID and flags indicating
whether music corresponds to a plurality of respective music
attributes are stored to be associated with each other, as shown in
FIG. 13. The music attribute may be, for example, a mood such as
relaxed, ballad, happy, active, and the like, and may be input as a
result of analysis processing by a computer.
[0041] The secondary list generation unit 80 reads flags associated
with each music ID included in a list stored in the primary list
generation unit 81 from the music attribute DB 82, and determines
whether or not there is music suitable for attributes included in a
request transmitted from the user device 12. Only music IDs of
pieces of music suitable for the attributes included in the request
transmitted from the user device 12 are extracted from music IDs
included in the list stored in the primary list storage unit 81,
and a list of the extracted music IDs is supplied to the sorting
unit 79 as a secondary list. The sorting unit 79 reads feature
vectors associated with the music IDs included in the secondary
list from the music feature DB 83, and reads a preference vector
stored in the preference vector storage unit 78 in association with
the user ID included in the request received from the user device
12. Similarity between two vectors is calculated and music IDs are
sorted in descending order of similarity. The similarity between
vectors may be an angle between two vectors. In this case, the
smaller the angle, the higher the similarity. The sorting unit 79
includes a predetermined number of (for example, 10) music IDs in a
recommendation music list in descending order of similarity between
two vectors, and returns the same list to the user device 12.
[0042] The music DB 85 is configured using the HDD 23 or a separate
DB (not shown), and stores data of music associated with a music
ID. The music distribution unit 84 receives a music ID from the
user device 12, reads music data stored in association with the
music ID from the music DB 85, and returns the music data to the
user device 12.
[0043] Here, an operation of the content recommendation system 10
according to this embodiment will be described. FIG. 14 is an
operation sequence diagram of the content recommendation system 10.
As shown in the same drawing, the user device 12 transmits a music
list request including a user ID and a music attribute to the music
distribution server 14 (S101). In the music distribution server 14,
the ranking selection and synthesis unit 75 reads user attributes
associated with the received user ID from the user information
storage unit 71 (S102), and acquires rankings corresponding to the
read user attributes from the internal ranking storage unit 74 and
the music ranking distribution server 15 (S103 and S104). The
acquired rankings are synthesized and stored in the primary list
storage unit 81 (S105). Next, in the music distribution server 14,
the secondary list generation unit 80 generates a secondary list by
narrowing down music IDs stored in the primary list storage unit 81
on the basis of music attributes (S106). Furthermore, the sorting
unit 79 sorts music IDs included in the secondary list according to
similarities between feature vectors associated with the music IDs
and a preference vector of a user (S107), and generates a
recommendation music list from a predetermined number of music IDs
in descending order of similarity to return the recommendation
music list to the user device 12 (S108). In the user device 12, the
music reproduction unit 62 receives the recommendation music list
and transmits the music IDs included in the recommendation music
list to the music distribution server 14 (S109). The music
distribution server 14 returns music data associated with the music
IDs received by the music distribution unit 84 to the user device
12 (S110). The music reproduction unit 62 of the user device 12
reproduces the received music data. Likewise, if music reproduction
for all the music IDs included in the recommendation music list is
performed, the user device 12 re-executes the process of S101.
[0044] Because a recommendation music list is generated using
various rankings, which are a list of content sequentially
generated over time, the above-described content recommendation
system 10 can prevent a problem of the same music being
continuously recommended to the user, and recommend various music
to the user.
[0045] The present invention is not limited to the above-described
embodiments, and various modifications are possible. Although an
example in which the music distribution server 14 generates a
recommendation music list has been described above, the functions
shown in FIG. 7, particularly, the ranking selection and synthesis
unit 75, the primary list storage unit 81, the secondary list
generation unit 80, the sorting unit 79, the preference vector
storage unit 78, and the preference vector update unit 77, are
provided in the user device 12, so that the user device 12 can
generate a recommendation music list.
* * * * *